Introduction

Top Abstract Introduction References

Nuclear factor-B (NF-B) is a protein transcription factor first identified by Sen and Baltimore (1) that functions to enhance the transcription of a variety of genes, including cytokines and growth factors, adhesion molecules, immunoreceptors, and acute-phase proteins. NF-B is required for maximal transcription of many cytokines (Table 1), including tumor necrosis factor- (TNF-), interleukin-1 (IL-1), IL-6, and IL-8, which are thought to be important in the generation of acute inflammatory responses. Excessive cytokine-mediated inflammation is likely to play a fundamental role in the pathogenesis of a variety of disease states, including sepsis syndrome and the adult respiratory distress syndrome (ARDS).

In general, cytokines are not stored intracellularly, and their secretion depends on new protein synthesis. As a consequence, elaboration of cytokines in response to an inflammatory stimulus is importantly or predominantly regulated by the transcription rates of cytokine genes. Since transcriptional regulation is critical for the production of many cytokines, transcription factors, including NF-B, may play key roles in regulating cytokine-mediated inflammation.

Current evidence suggests that cytokines function in redundant and overlapping ways through so-called cytokine "cascades" or "networks." Although NF-B appears to play a critical role in cytokine-mediated inflammation by upregulating the transcription of a specific set of cytokine genes in response to inflammatory stimuli, it is uncertain whether NF-B has a significant role in the differential production of NF-B-dependent cytokines or in coordinating the production of these cytokines. The timing of cytokine production and relative amount of cytokines produced by a stimulus are probably functions of interactions between NF-B and other transcription factors, as well as factors independent of NF-B.

Organ-system dysfunction in a variety of inflammatory diseases appears to be determined either directly or indirectly by an overproduction of cytokine-mediated inflammation. For the purpose of intervening therapeutically in these diseases and modulating the entire cytokine network, it would be valuable to exploit mechanisms common to the production of many cytokines, such as transcriptional regulation by NF-B. Therefore, understanding of the function of NF-B and other transcription factors may be fundamental to the study of cytokines and cytokine-mediated inflammation, and may provide novel therapeutic strategies for a number of inflammatory diseases.

In this review, we will succinctly discuss the molecular biology of the Rel family of proteins, which includes NF-B. Recent, comprehensive reviews have detailed the structure, functions, and interactions of this protein family (2- 4). We will then examine the role of NF-B in regulating cytokine production, explore the intricate system of positive and negative feedback loops that control NF-B activation, and evaluate current information about the significance of NF-B in the pathobiology of disease. Additionally, we will discuss strategies for modulating NF-B acti-vation and potentially alter cytokine-mediated inflammatory responses. Our review will focus primarily on NF-B regulation of cytokine production in leukocytes. Although NF-B activation and cytokine production occur in a variety of cell types, most available information about the activation and regulation of NF-B in relationship to cytokine production has been derived from studies involving lymphocytes, monocytes, or macrophages. Regulation of NF-B activation and its effect on cytokine production may be different in nonimmune cells than in leukocytes, but this is not currently well understood.

	Molecular Biology of NF-B

Top Abstract Introduction References

NF-B consists of two members of the Rel family of proteins. As with other transcription factors, NF-B attaches to DNA in the promoter regions of target genes as a dimer composed of two Rel family proteins, p50 (NF-B1) and RelA (p65). In the NF-B heterodimer, both subunits contact DNA, but only RelA contains a transactivation domain in the C-terminal end of the protein that activates transcription by direct interaction with the basal transcription apparatus (5). The Rel family contains other members, including c-Rel, RelB, and p52 (NF-B2), which in combination with p50 and RelA exist in a wide variety of cell types and can form various hetero- and homodimers (Table 2). Although NF-B is classically defined as a p50/ RelA heterodimer, other combinations of Rel proteins can function identically to NF-B. For this reason, we will refer to any combination of Rel proteins that bind to NF-B-binding sites as NF-B.

In quiescent cells, NF-B is sequestered in the cytoplasm through its interaction with the inhibitors IB-, IB-, or IB- (Table 2). In addition, p105, which is the precursor of p50, and p100, which is the precursor of p52, can bind RelA and thus function as NK-B inhibitors. The C-terminal portions of p105 and p100 have been designated IB- and IB-, respectively. These inhibitory units contain ankyrin repeat domains that allow interaction with NF-B in a configuration that masks nuclear localization signal domains, thereby preventing nuclear transport (3, 4). IB- interacts only weakly with p50 homodimers, and does not efficiently prevent their translocation to the nucleus (4).

NF-B can be activated in cells by a variety of stimuli, including bacterial endotoxin, TNF-, IL-1, mitogens, viral proteins, ionizing radiation, UV light, and certain chemical agents (4). Following activation, the inhibitory units are phosphorylated and degraded, unmasking nuclear localization signals that allow NF-B to be transported to the cell nucleus, where its dimers are free to bind DNA containing the sequence (5'-GGGPuNNPyPyCC-3'). The mechanism of processing of the inhibitory units has been a subject of much recent investigation, and is currently best understood for IB-. Initially, IB- (bound to NF-B) is phosphorylated at ser³² and ser³⁶, subsequent to a signal originating at the cell surface (6). Recent evidence shows that this phosphorylation can be effected by a specific IB kinase that is dependent on ubiquitin (7). Other studies suggest that several different kinases have the potential to phosphorylate IB-. Phosphorylation of IB- serves as a tag for the addition of ubiquitin (8), which leads to recognition of the IB- molecule by the proteasome complex and subsequent degradation of the IB- molecule, freeing NF-B to translocate to the nucleus.

The Rel protein designated p105 is the precursor of the p50 subunit of NF-B, but also functions as an inhibitor by binding RelA and retaining it in the cytoplasm. NF-B is activated by proteolytic cleavage and degradation of the C-terminal fragment of p105. Recent evidence suggests that the ubiquitin-proteasome system is involved in the proteolytic processing of p105 (9) as well as IB-.

The IB family contains other members, including IB-, IB-, and Bcl-3. IB- binds to RelA and c-Rel, but not to p50, and inhibits movement of these proteins to the nucleus. The regulation of IB- processing is not currently well understood. IB- is another member of the IB family that has been recently described (10). This inhibitor also binds to and inhibits RelA- and c-Rel-containing complexes. Bcl-3 is a unique protein that can be present in the nucleus and has the ability to bind p50 and p52 homodimers in certain cells and to function as a transactivator. The relative importance of these different inhibitors in regulating NF-B activation is uncertain, but the presence of multiple inhibitors is clearly important in the intricate system of checks and balances that controls NF-B activation.

	NF-B Regulation of Cytokine Networks

Top Abstract Introduction References

Although cytokines may act independently, inflammation is usually associated with their coordinated production and action. In sepsis syndrome, for example, bacterial endotoxin and other toxic products stimulate rapid production of TNF- and IL-1 by a variety of host cells. These cytokines mediate many of the early host responses to infection, and can cause macrophages and other cell types to secrete additional cytokines, such as IL-6 and IL-8, which have profound consequences on the host. NF-B exerts a broad influence over this network of cytokines by affecting transcription of many of the genes involved in its generation. Cytokine networks are not limited to sepsis, and may also be important in the pathobiology of other inflammatory diseases such as asthma and rheumatoid arthritis.

Following a stimulus, cytokines are produced in a characteristic pattern that is dependent on the stimulus as well as the cell or tissue type. Although NF-B is known to function as an activator of transcription, its role in the differential transcription of NF-B-dependent cytokines is less well defined. Certainly interactions between NF-B and other activated transcription factors are important for determining the transcription rates of cytokine genes. NF-B may also produce preferential binding and transactivation at certain NF-B motifs, which could favor the production of certain cytokines.

NF-B interacts with the basal transcription apparatus as well as many other transactivators and repressors in the context of each individual promoter to coordinate transcription. Interactions with other promoter-bound transcription factors are crucial for regulating the expression of cytokine genes. For example, cooperative binding with the transcription factor NF-IL-6 is required for the transcriptional activation of IL-8 and IL-6 (11, 12). In addition, NF-B may have direct protein-protein interactions with other transcription factors, such as the glucocorticoid receptor, that alter the ability of NF-B to bind to DNA (13).

NF-B and other Rel proteins bind to similar sites but with different affinities. Kunsch and colleagues have shown that different NF-B motifs have different affinities for different Rel protein dimers (14). Also, Lin and coworkers have shown that different Rel dimers have different abilities to stimulate transcription when bound to the same NF-B motif (15). The implication of these studies is that different Rel dimers may preferentially bind and transactivate at certain NF-B motifs.

Several other factors are critical in determining the pattern of cytokine production following an inflammatory stimulus. Factors such as the rate of posttranscriptional RNA processing, mRNA stability, and translation efficiency vary considerably among cytokines. For TNF-, posttranscriptional events are relatively more important than transcriptional activation in determining the quantity of TNF- produced. When macrophages are stimulated by endotoxin to produce TNF-, the transcription rates for this cytokine increase by 5- to 50-fold, but the efficiency of translation increases more than 100-fold (16). In contrast, endotoxin-induced IL-8 production in macrophages depends primarily on increased IL-8 gene transcription and RNA processing (17). In addition, there are cell-specific differences that lead to the production of a specific set of cytokines. For instance, differences in chromatin may make some promoters more accessible than others in certain cells. Also, cells have varying complements of transcription factors available for activation, depending on the cell type and activation state. In summary, the specificity and timing of cytokine production in response to a given stimulus are probably determined by a complex interaction of NF-B with a variety of NF-B binding sites and an array of other transcription factors, as well as factors independent of NF-B.

	Positive and Negative Feedback Loops in Regulation of NF-B Activation

Top Abstract Introduction References

Because NF-B is an integral and critical regulator of cytokine-mediated inflammation, the activation of NF-B is a tightly controlled event. Feedback control of NF-B activation occurs both intracellularly and extracellularly (Figure 1).

View larger version (32K): [in this window] [in a new window]   Figure 1.   Regulation of NF-B (RelA/p50) activation following cell stimulation with TNF-, IL-1, or endotoxin (LPS). Binding of these agents at the cell surface signals phosphorylation and addition of ubiquitin (Ub) to the inhibitory unit (IB- or p105). These events lead to recognition and proteolysis of the inhibitor, exposing the nuclear localization signal (NLS) on p50 and RelA. RelA/p50 heterodimers bind to specific motifs in the promoter regions of cytokine genes, and initiate transcription. Cytokine mRNA is translated into proteins, which are secreted by cells. TNF- and IL-1 can act locally to amplify NF-B activation. Nuclear binding of NF-B also stimulates production of IB- and p105, which inhibit further activation of NF-B. Increased p105 production favors the formation of p50 homodimers, which can inhibit NF-B action by entering the nucleus and competing for NF-B binding sites.

Positive feedback may occur through extracellular mechanisms that serve to amplify inflammatory signals. NF-B activation enhances the transcription of TNF- and IL-1, and both of these cytokines are in turn known to activate NF-B. An inflammatory signal, such a bacterial endotoxin, can cause cells to activate NF-B, which enhances TNF- and IL-1 production and release, and presumably could amplify the original inflammatory signal. This mechanism may occur in sepsis syndrome, in which TNF- and IL-1 are released into serum early in the course of the disorder. Other mediators, such as IL-6 and IL-8, are released later and have more sustained elevations. These later mediators may depend largely on TNF- and IL-1 to stimulate their production.

Negative feedback control is essential in regulating NF-B activation. Both intracellular and extracellular mechanisms are responsible for limiting NF-B activation in response to a given stimulus. Intracellularly, NF-B activation leads to transcriptional upregulation of the IB- and p105 genes, since both of these genes have NF-B-responsive elements in their promoters (18, 19). Increased production of inhibitory units presumably helps trap NF-B in the cytoplasmic compartment, and downregulates activated nuclear NF-B, thus terminating new cytokine transcription and limiting the inflammatory response. An interesting effect of increased production of p105 is that p50 homodimer formation is also increased, which may diminish NF-B-mediated responses to subsequent stimuli. Since p50 homodimers do not bind efficiently to IB, and lack transcription-activation domains, they can translocate to the nucleus and function as inhibitors of NF-B-mediated gene expression by competing with other Rel proteins for access to NF-B binding sites. Zeigler-Heitbrock and colleagues have demonstrated increased p50 homodimer production in cell models of endotoxin tolerance (20).

In addition to intracellular feedback control of NF-B activation, there appear to be extracellular mechanisms for limiting inflammatory responses through downregulation of NF-B. Inflammatory stimuli such as endotoxin, TNF-, and IL-1 can stimulate the production of counterregulatory cytokines, such as IL-10, that suppress the production of proinflammatory cytokines. Wang and associates have recently shown that IL-10 can inhibit cytokine production in monocytes by blocking endotoxin-induced NF-B activation, although the mechanism for this effect is unknown (21). In all, these findings provide evidence for an intricate feedback control of NF-B activation that involves extracellular feedback mechanisms that both stimulate and suppress NF-B activation, and intracellular feedback mechanisms that limit NF-B activation to a given stimulus.

	NF-B in Human Disease and Animal Models

Top Abstract Introduction References

Although the importance of NF-B in cytokine transcription has been established in vitro, investigations into the role of NF-B in human disease have only recently been undertaken. Schwartz and coworkers (22) have reported that NF-B in alveolar macrophages from patients with ARDS is activated to a significantly higher degree than in alveolar macrophages from critically ill patients with other diseases. This finding correlates with previous reports that IL-8 and TNF- are increased in lung lavage samples from patients with ARDS (23, 24). In addition, Asahara and colleagues (25) recently showed that NF-B is activated in the synovium of patients with rheumatoid arthritis as compared with patients with osteoarthritis. Currently, experimental evidence linking NF-B activation in specific cells or tissues to other inflammatory diseases in humans is lacking.

Several animal models have been developed to evaluate the role of NF-B in the production of inflammatory events. We have described a rat model of neutrophilic lung inflammation following intraperitoneal endotoxin injection (26). In this model, endotoxin injection is followed by activation of NF-B in alveolar macrophages and in lung tissue (26, 27). Activation of NF-B correlates with expression of mRNA for cytokine-induced neutrophil chemoattractant (CINC), a neutrophil chemotactic chemokine, and these events are followed by an influx of neutrophils into the alveolar space. In addition, we have recently shown that blocking endotoxin-induced NF-B activation in lung tissue results in decreased CINC mRNA expression and diminished neutrophilic lung inflammation (27). This finding supports the concept that regulating NF-B activation can alter inflammatory events. Others have shown that NF-B activation in lung tissue following an inflammatory stimulus correlates with cytokine gene expression. Shenkar and colleagues (28) showed that NF-B is activated in mouse lung tissue by hypovolemic shock, which also leads to the activation of cytokine production. In addition, Haddad and associates (29) demonstrated that ozone exposure induces NF-B activation and CINC mRNA expression in rat lungs, and that this can be blocked by treatment with corticosteroids.

In animal models of disease, NF-B activation has been reported at sites of inflammation other than the lung. For example, NF-B activation has been shown in rat microglial cells in a model of autoimmune encephalomyelitis (30). Also, Neurath and coworkers (31) recently reported an interesting study in which experimental colitis in mice was effectively blocked by the administration of antisense oligonucleotides to the RelA subunit of NF-B. In a rat model of glomerulonephritis, NF-B activation has been shown in glomeruli (32). Blocking of NF-B activation in this model, by treatment with pyrrolidine dithiocarbamate, led to decreased glomerular cytokine mRNA expression and diminished urinary protein excretion. Together, these studies link in vivo NF-B activation with cytokine production and the generation of inflammation.

	Studies with NF-B/Rel and IB Knockout Mice

Top Abstract Introduction References

In order to better understand the roles of specific Rel family members in vivo, knockout mice have been created for RelA, IB-, p50, c-Rel, and RelB (33). Studies done with these knockout mice as well as a recent study by Schmidt-Ulrich and coworkers (39), have shed light on the role of Rel family proteins in embryonic development. Schmidt-Ulrich and coworkers evaluated the role of NF-B in development by using transgenic mice with NF-B-driven LacZ reporter constructs. In contrast to Rel family protein Dorsal, which is activated early in Drosophila embryogenesis, NF-B activation was not detected early in mouse development. Deficiencies of p50, c-Rel, or RelB result in developmentally normal mice, but RelA deficiency results in embryonic lethality due to liver apoptosis (33). On the basis of these observations, NF-B appears to be more important in maintaining organ function than in early development or tissue differentiation. There is some evidence, however, that c-Rel is important in limb-bud development in mice (10).

Studies involving NF-B/Rel and IB knockout mice have demonstrated the pivotal role of these transcription factors in immune-system function. IB--deficient mice are apparently normal at birth, but postnatally, their growth ceases and they die by 7 to 10 days of age (34, 35). Death in these animals is reported to occur in association with enhanced granulopoiesis, severe dermatitis, and increased TNF- in the skin. Splenocytes from these animals demonstrate increased NF-B activation, whereas fibroblasts show minimal spontaneous NF-B activation; however, treatment with TNF- of fibroblasts from these mice results in prolonged activation of NF-B as compared with fibroblasts from normal mice (34, 35). The finding of prolonged upregulation of NF-B activation in stimulated fibroblasts confirms the role of IB- in limiting NF-B activation.

Mice deficient in p50 have defects in B-lymphocyte function and altered susceptibility to infection (36). These mice show defective clearance of Listeria monocytogenes and Streptococcus pneumoniae, but greater resistance to infection with murine encephalomyocarditis virus. When stimulated with endotoxin, peritoneal macrophages from these animals exhibited normal TNF- and IL-1 release and decreased IL-6 release as compared with normal mouse macrophages (36). On the basis of these observations, it appears that p50 is critical for mediating certain immune responses. c-Rel-deficient mice have impaired B- and T-lymphocyte function (37). Disruption of RelB in mice leads to phenotypic changes including multiorgan inflammation involving the liver and lung, among other organs, as well as to impaired cellular immunity, splenomegaly, and myeloid hyperplasia in bone marrow (38). In combination, these studies show that deletion of specific Rel-family genes in mice leads to multiple immune defects; however, the full impact of deficiencies of specific Rel proteins may be masked by the redundancy of the NF-B/Rel protein family. Nonetheless, these findings illustrate the critical and complex interplay of this family of transcription-regulating proteins in the normally functioning immune system.

	Modulating NF-B Activation and Modifying the Inflammatory Response

Top Abstract Introduction References

Because of its potential ability to influence the production of an array of cytokines, NF-B is an appealing target for therapeutic strategies designed to attenuate cytokine-mediated inflammation. A number of compounds have been identified that can suppress NF-B activation in vitro, including antioxidants, protease inhibitors, proteasome inhibitors, corticosteroids, salicylates, and other immunosuppressants.

In terms of NF-B inhibition, antioxidants are the best-studied class of agents. Antioxidants have been investigated as inhibitors of NF-B activation because the generation of reactive oxygen species (ROS) is postulated to be a vital link in mediating NF-B activation by a variety of stimuli. Four lines of evidence support this concept of ROS as playing a role in NF-B activation. First, direct treatment with oxidants such as H₂O₂ activates NF-B in some cells (40). Second, agents that activate NF-B in cells (including endotoxin, TNF-, IL-1, and ionizing radiation) produce oxidative stress (40). Systemic endotoxin treatment in rats induces both oxidative stress and NF-B activation in lung tissue (27). As a third line of evidence for a link between ROS and NF-B, antioxidants have been shown to inhibit NF-B activation in a variety of settings both in vitro (40) and in vivo. We have shown that the antioxidant N-acetylcysteine (NAC) can inhibit NF-B activation in lung tissue following systemic endotoxin treatment (27). Moreover, upregulation of endogenous oxidant defenses has been shown to suppress NF-B activation. Mirochnitchenko and Inouye (41) have shown that peritoneal macrophages from transgenic mice that overexpress human Cu, Zn superoxide dismutase (CuZn SOD) have decreased NF-B activation in response to stimulation with phorbol 12-myristate 13-acetate (PMA) as compared with peritoneal macrophages from control mice. Others have shown that cell lines that overproduce catalase but not SOD have a diminished ability to activate NF-B (42). Since ROS appear to be important intermediates in NF-B activation, inhibiting their generation or effect might be beneficial in limiting inflammation in certain clinical settings.

Corticosteroids, a group of compounds with a broad range of effects on the immune system, appear to block NF-B activation in two ways. First, glucocorticoid receptors can interact directly with RelA to inhibit DNA binding (13). Second, corticosteroids activate the production of IB-, which downregulates NF-B (43, 44). Corticosteroids, like all pharmacologic agents that have been shown to inhibit NF-B, have numerous other effects that could limit their therapeutic usefulness. Currently, there is much interest in identifying more specific and effective NF-B inhibitors.

We have recently investigated a different approach to downregulating NF-B-dependent cytokine production. We attempted to make rats endotoxin tolerant by giving them four daily injections of endotoxin. When endotoxin-tolerant rats are subsequently treated with high-dose intraperitoneal injections of endotoxin, they have decreased NF-B activation and chemokine gene expression in lung tissue, as well as attenuated neutrophilic lung inflammation, as compared with endotoxin-sensitive rats (T. S. Blackwell and J. W. Christman: unpublished data). Exploiting the natural phenomenon of endotoxin tolerance may be an effective way to suppress NF-B-dependent inflammation.

	Future Directions

Top Abstract Introduction References

In the 10 yr since the first publication by Sen and Baltimore describing NF-B, investigation of transcription regulation by NF-B has burgeoned. NF-B has been shown to be a tightly regulated agent for initiating the transcription of a wide variety of genes involved in the production of acute inflammation; however, several specific issues need to be addressed in future studies. Further investigation is needed into the signal-transduction process that leads to the degradation of IB-family proteins and the activation of NF-B following an inflammatory stimulus. In addition, the cell- and gene-specific action and regulation of NF-B need to be further explored. Comparing the regulation and effects of NF-B activation in nonimmune cells with those in immune cells is likely to yield important information. Another important area of ongoing research is the investigation of interactions between NF-B and other transcription factors in regulating the differential production of cytokines in specific cell types.

Research on the role of transcription factors in inflammatory diseases is in its very early stages. Further investigation is warranted to determine whether the intensity of NF-B activation is useful as a marker for the severity of inflammation in certain diseases, and whether NF-B activation could be useful as a surrogate marker for assessing the efficacy of therapeutic interventions. Specific inhibitors of NF-B would be beneficial in further dissecting the role of NF-B in the complex acute inflammatory response, and could be clinically useful in treating inflammatory diseases.